DEI - Conference Room
April 26th, 2010
I will briefly define the general notion of Recommender System, to show that these systems are in fact examples of more general information search and discovery tools.
I will try also to classify these tools. While information seeking is normally supported by search engines, information discovery is much more difficult and there are many user tasks that can be included in this general concept.
I will present some interfaces for supporting information discovery, including interfaces for recommender systems.
I will talk about conversational systems, since they offer a more flexible and involving approach to recommendations.
Then I will talk about issues related to the evaluation of these systems, contrasting the off-line methods with online ones. I will stress some problems related to the evaluation of recommender systems in both scenarios, presenting my model of the evaluation and the range of effects of a recommendation on the user.
Then I will briefly mention some psychological issues that we must take into account when building interface for recommender systems and how this can impact on the design of the user interface and on the evaluation methodology.